Análisis de modelos SOM - Frecuencia datos de entrada: mes

Modelo

• ID: 323
• Descripción:
• Frecuencia: mes
• Variables: fecha_cnt, tmax, tmin, precip, nevada, prof_nieve
• Dimensiones del mapa: 10,10
• Iteraciones: 1000
• Parámetros adicionales:

source("../../lib/som-utils.R")

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

source("../../lib/maps-utils.R")

Linking to GEOS 3.8.0, GDAL 3.0.4, PROJ 6.3.1

Carga del modelo desde disco

mpr.set_base_path_analysis()
model <- mpr.load_model("som-323.rds.xz")
summary(model)

SOM of size 10x10 with a hexagonal topology and a bubble neighbourhood function.
The number of data layers is 1.
Distance measure(s) used: sumofsquares.
Training data included: 94881 objects.
Mean distance to the closest unit in the map: 0.425.

plot(model, type="changes")

Carga del dataset de entrada

df <- mpr.load_data("datos_mes.csv.xz")

df

summary(df)

 id_estacion           fecha             fecha_cnt           tmax      
 Length:94881       Length:94881       Min.   : 1.000   Min.   :-53.0  
 Class :character   Class :character   1st Qu.: 4.000   1st Qu.:148.0  
 Mode  :character   Mode  :character   Median : 6.000   Median :198.0  
                                       Mean   : 6.497   Mean   :200.2  
                                       3rd Qu.: 9.000   3rd Qu.:255.0  
                                       Max.   :12.000   Max.   :403.0  
      tmin             precip           nevada           prof_nieve      
 Min.   :-121.00   Min.   :  0.00   Min.   :0.000000   Min.   :   0.000  
 1st Qu.:  53.00   1st Qu.:  3.00   1st Qu.:0.000000   1st Qu.:   0.000  
 Median :  98.00   Median : 10.00   Median :0.000000   Median :   0.000  
 Mean   :  98.86   Mean   : 16.25   Mean   :0.000295   Mean   :   0.467  
 3rd Qu.: 148.00   3rd Qu.: 22.00   3rd Qu.:0.000000   3rd Qu.:   0.000  
 Max.   : 254.00   Max.   :422.00   Max.   :6.000000   Max.   :1834.000  
    longitud        latitud            altitud      
 Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.:38.28   1st Qu.: -5.6417   1st Qu.:  42.0  
 Median :40.82   Median : -3.4500   Median : 247.0  
 Mean   :39.66   Mean   : -3.4350   Mean   : 418.5  
 3rd Qu.:42.08   3rd Qu.:  0.4914   3rd Qu.: 656.0  
 Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

Carga de los mapas

world <- ne_countries(scale = "medium", returnclass = "sf")
spain <- subset(world, admin == "Spain")

Mapa de densidad

plot(model, type="count", shape = "straight", palette.name = mpr.degrade.bleu)

Número de elementos en cada celda:

nb <- table(model$unit.classif)
print(nb)

   1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16 
 547 1079 1482 1222 1865 1818 1254 1715 1698 1794 1458 1159 1098 1261 1546 1157 
  17   18   19   20   21   22   23   24   25   26   27   28   29   30   31   32 
1669 1543 1587 1484 1521 1594 1370 1401 1408  774 1262 1240 1105  760 1559 1257 
  33   34   35   36   37   38   39   40   41   42   43   44   45   46   47   48 
 597 1464 1352  884  992 1179 1297 1350 1147 1153 1616 1071  860  665  977 1260 
  49   50   51   52   53   54   55   56   57   58   59   60   61   62   63   64 
 777 1426 1039  930  320 1364 1265  660 1018 1057  928 1452 1412 1417 1090  884 
  65   66   67   68   69   70   71   72   73   74   75   76   77   78   79   80 
 605  450  809  717 1105 1646 1043   74  836  486  777  185  418  495 1036    3 
  81   82   83   84   85   86   87   88   89   90   91   92   93   94   95   96 
  19   39  298  460  352  203  328  493 1118 1049    5  251  139  278   93   26 
  97   98   99  100 
  78  164  204    9 

Comprobación de nodos vacíos:

dim_model <- 10*10;
len_nb = length(nb);
empty_nodes <- dim_model != len_nb;
if (empty_nodes) {
  print(paste("[Warning] Existen nodos vacíos: ", len_nb, "/", dim_model))
}

Mapa de distancia entre vecinos

plot(model, type="dist.neighbours", shape = "straight")

Influencia de las variables

model_colnames = c("fecha_cnt", "tmax", "tmin", "precip", "nevada", "prof_nieve")
model_ncol = length(model_colnames)

Mapa de variables.

plot(model, shape = "straight")

Mapa de calor por variable

par(mfrow=c(3,4))
for (j in 1:model_ncol) {
  plot(model, type="property", property=getCodes(model,1)[,j],
    palette.name=mpr.coolBlueHotRed,
    main=model_colnames[j],
    cex=0.5, shape = "straight")
}

Correlación para cada columna del vector de nodos

if (!empty_nodes) {
  cor <- apply(getCodes(model,1), 2, mpr.weighted.correlation, w=nb, som=model)
  print(cor)
}

      fecha_cnt       tmax       tmin      precip     nevada  prof_nieve
[1,] 0.91407932  0.3201151  0.3690164 -0.03192877 0.01480024 -0.03949785
[2,] 0.07691277 -0.7672002 -0.7197780  0.64363659 0.02387252  0.05783297

Representación de cada variable en un mapa de factores:

if (!empty_nodes) {
  par(mfrow=c(1,1))
  plot(cor[1,], cor[2,], xlim=c(-1,1), ylim=c(-1,1), type="n")
  lines(c(-1,1),c(0,0))
  lines(c(0,0),c(-1,1))
  text(cor[1,], cor[2,], labels=model_colnames, cex=0.75)
  symbols(0,0,circles=1,inches=F,add=T)
}

Importancia de cada variable - varianza ponderada por el tamaño de la celda:

if (!empty_nodes) {
  sigma2 <- sqrt(apply(getCodes(model,1),2,function(x,effectif)
     {m<-sum(effectif*(x-weighted.mean(x,effectif))^2)/(sum(effectif)-1)},
     effectif=nb))
  print(sort(sigma2,decreasing=T))
}

 fecha_cnt       tmin       tmax     precip prof_nieve     nevada 
 0.9810094  0.9756455  0.9754580  0.9749566  0.9685620  0.9034467 

Clustering

if (!empty_nodes) {
  hac <- mpr.hac(model, nb)
}

Visualización de 3 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=3)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=3)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin             precip      
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.00   Min.   :  0.00  
 1st Qu.: 4.000   1st Qu.:148.0   1st Qu.:  53.00   1st Qu.:  3.00  
 Median : 6.000   Median :198.0   Median :  98.00   Median : 10.00  
 Mean   : 6.497   Mean   :200.2   Mean   :  98.87   Mean   : 16.25  
 3rd Qu.: 9.000   3rd Qu.:255.0   3rd Qu.: 148.00   3rd Qu.: 22.00  
 Max.   :12.000   Max.   :403.0   Max.   : 254.00   Max.   :422.00  
     nevada           prof_nieve          longitud        latitud        
 Min.   :0.00e+00   Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889  
 1st Qu.:0.00e+00   1st Qu.:  0.0000   1st Qu.:38.28   1st Qu.: -5.6417  
 Median :0.00e+00   Median :  0.0000   Median :40.82   Median : -3.4500  
 Mean   :3.16e-05   Mean   :  0.3973   Mean   :39.66   Mean   : -3.4351  
 3rd Qu.:0.00e+00   3rd Qu.:  0.0000   3rd Qu.:42.08   3rd Qu.:  0.4914  
 Max.   :1.00e+00   Max.   :892.0000   Max.   :43.57   Max.   :  4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  42.0  
 Median : 247.0  
 Mean   : 418.5  
 3rd Qu.: 656.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt        tmax           tmin           precip         nevada 
 Min.   :2.0   Min.   :-4.0   Min.   :-51.0   Min.   :19.0   Min.   :0  
 1st Qu.:2.0   1st Qu.: 0.0   1st Qu.:-48.0   1st Qu.:33.0   1st Qu.:0  
 Median :3.0   Median :35.0   Median :-26.0   Median :49.0   Median :0  
 Mean   :2.6   Mean   :20.6   Mean   :-32.8   Mean   :41.4   Mean   :0  
 3rd Qu.:3.0   3rd Qu.:36.0   3rd Qu.:-23.0   3rd Qu.:50.0   3rd Qu.:0  
 Max.   :3.0   Max.   :36.0   Max.   :-16.0   Max.   :56.0   Max.   :0  
   prof_nieve      longitud        latitud         altitud    
 Min.   :1017   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:1073   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :1168   Median :40.78   Median :-4.01   Median :1894  
 Mean   :1317   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:1494   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

Visualización de 4 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=4)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=4)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin            precip      
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.0   Min.   :  0.00  
 1st Qu.: 4.000   1st Qu.:148.0   1st Qu.:  53.0   1st Qu.:  3.00  
 Median : 6.000   Median :198.0   Median :  98.0   Median : 10.00  
 Mean   : 6.498   Mean   :200.3   Mean   :  98.9   Mean   : 16.24  
 3rd Qu.: 9.000   3rd Qu.:255.0   3rd Qu.: 148.0   3rd Qu.: 22.00  
 Max.   :12.000   Max.   :403.0   Max.   : 254.0   Max.   :422.00  
     nevada           prof_nieve          longitud        latitud        
 Min.   :0.00e+00   Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889  
 1st Qu.:0.00e+00   1st Qu.:  0.0000   1st Qu.:38.28   1st Qu.: -5.6417  
 Median :0.00e+00   Median :  0.0000   Median :40.82   Median : -3.4500  
 Mean   :3.16e-05   Mean   :  0.2746   Mean   :39.66   Mean   : -3.4350  
 3rd Qu.:0.00e+00   3rd Qu.:  0.0000   3rd Qu.:42.08   3rd Qu.:  0.4914  
 Max.   :1.00e+00   Max.   :429.0000   Max.   :43.57   Max.   :  4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  42.0  
 Median : 247.0  
 Mean   : 418.2  
 3rd Qu.: 656.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt          tmax             tmin            precip           nevada 
 Min.   :1.000   Min.   :-12.00   Min.   :-72.00   Min.   : 19.00   Min.   :0  
 1st Qu.:2.000   1st Qu.:  9.50   1st Qu.:-43.50   1st Qu.: 41.00   1st Qu.:0  
 Median :2.000   Median : 15.00   Median :-34.00   Median : 56.00   Median :0  
 Mean   :2.368   Mean   : 23.21   Mean   :-31.95   Mean   : 65.79   Mean   :0  
 3rd Qu.:3.000   3rd Qu.: 28.50   3rd Qu.:-25.00   3rd Qu.: 82.00   3rd Qu.:0  
 Max.   :4.000   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :465.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:531.0   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :606.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :612.8   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:661.5   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :892.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt        tmax           tmin           precip         nevada 
 Min.   :2.0   Min.   :-4.0   Min.   :-51.0   Min.   :19.0   Min.   :0  
 1st Qu.:2.0   1st Qu.: 0.0   1st Qu.:-48.0   1st Qu.:33.0   1st Qu.:0  
 Median :3.0   Median :35.0   Median :-26.0   Median :49.0   Median :0  
 Mean   :2.6   Mean   :20.6   Mean   :-32.8   Mean   :41.4   Mean   :0  
 3rd Qu.:3.0   3rd Qu.:36.0   3rd Qu.:-23.0   3rd Qu.:50.0   3rd Qu.:0  
 Max.   :3.0   Max.   :36.0   Max.   :-16.0   Max.   :56.0   Max.   :0  
   prof_nieve      longitud        latitud         altitud    
 Min.   :1017   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:1073   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :1168   Median :40.78   Median :-4.01   Median :1894  
 Mean   :1317   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:1494   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

Visualización de 5 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=5)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=5)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin            precip           nevada 
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.0   Min.   :  0.00   Min.   :0  
 1st Qu.: 4.000   1st Qu.:149.0   1st Qu.:  53.0   1st Qu.:  3.00   1st Qu.:0  
 Median : 6.000   Median :198.0   Median :  98.0   Median : 10.00   Median :0  
 Mean   : 6.498   Mean   :200.3   Mean   :  98.9   Mean   : 16.24   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:255.0   3rd Qu.: 148.0   3rd Qu.: 22.00   3rd Qu.:0  
 Max.   :12.000   Max.   :403.0   Max.   : 254.0   Max.   :422.00   Max.   :0  
   prof_nieve          longitud        latitud            altitud      
 Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.:  0.0000   1st Qu.:38.28   1st Qu.: -5.6417   1st Qu.:  42.0  
 Median :  0.0000   Median :40.82   Median : -3.4500   Median : 247.0  
 Mean   :  0.2746   Mean   :39.66   Mean   : -3.4350   Mean   : 418.2  
 3rd Qu.:  0.0000   3rd Qu.:42.08   3rd Qu.:  0.4914   3rd Qu.: 656.0  
 Max.   :429.0000   Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax            tmin           precip          nevada 
 Min.   : 1.000   Min.   :86.00   Min.   :-24.0   Min.   : 6.00   Min.   :1  
 1st Qu.: 6.500   1st Qu.:86.00   1st Qu.:-19.5   1st Qu.: 6.00   1st Qu.:1  
 Median :12.000   Median :86.00   Median :-15.0   Median : 6.00   Median :1  
 Mean   : 8.333   Mean   :88.33   Mean   : -7.0   Mean   :10.33   Mean   :1  
 3rd Qu.:12.000   3rd Qu.:89.50   3rd Qu.:  1.5   3rd Qu.:12.50   3rd Qu.:1  
 Max.   :12.000   Max.   :93.00   Max.   : 18.0   Max.   :19.00   Max.   :1  
   prof_nieve        longitud        latitud          altitud     
 Min.   :0.0000   Min.   :40.48   Min.   :-3.450   Min.   :263.0  
 1st Qu.:0.0000   1st Qu.:40.48   1st Qu.:-3.450   1st Qu.:435.6  
 Median :0.0000   Median :40.48   Median :-3.450   Median :608.1  
 Mean   :0.3333   Mean   :40.88   Mean   :-2.644   Mean   :493.1  
 3rd Qu.:0.5000   3rd Qu.:41.08   3rd Qu.:-2.242   3rd Qu.:608.1  
 Max.   :1.0000   Max.   :41.67   Max.   :-1.033   Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt          tmax             tmin            precip           nevada 
 Min.   :1.000   Min.   :-12.00   Min.   :-72.00   Min.   : 19.00   Min.   :0  
 1st Qu.:2.000   1st Qu.:  9.50   1st Qu.:-43.50   1st Qu.: 41.00   1st Qu.:0  
 Median :2.000   Median : 15.00   Median :-34.00   Median : 56.00   Median :0  
 Mean   :2.368   Mean   : 23.21   Mean   :-31.95   Mean   : 65.79   Mean   :0  
 3rd Qu.:3.000   3rd Qu.: 28.50   3rd Qu.:-25.00   3rd Qu.: 82.00   3rd Qu.:0  
 Max.   :4.000   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :465.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:531.0   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :606.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :612.8   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:661.5   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :892.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt        tmax           tmin           precip         nevada 
 Min.   :2.0   Min.   :-4.0   Min.   :-51.0   Min.   :19.0   Min.   :0  
 1st Qu.:2.0   1st Qu.: 0.0   1st Qu.:-48.0   1st Qu.:33.0   1st Qu.:0  
 Median :3.0   Median :35.0   Median :-26.0   Median :49.0   Median :0  
 Mean   :2.6   Mean   :20.6   Mean   :-32.8   Mean   :41.4   Mean   :0  
 3rd Qu.:3.0   3rd Qu.:36.0   3rd Qu.:-23.0   3rd Qu.:50.0   3rd Qu.:0  
 Max.   :3.0   Max.   :36.0   Max.   :-16.0   Max.   :56.0   Max.   :0  
   prof_nieve      longitud        latitud         altitud    
 Min.   :1017   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:1073   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :1168   Median :40.78   Median :-4.01   Median :1894  
 Mean   :1317   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:1494   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

Visualización de 6 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=6)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=6)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)
  df.cluster06 <- subset(df, cluster==6)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster06 <- select(df.cluster06, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin             precip      
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.00   Min.   :  0.00  
 1st Qu.: 4.000   1st Qu.:149.0   1st Qu.:  53.00   1st Qu.:  3.00  
 Median : 6.000   Median :198.0   Median :  98.00   Median : 10.00  
 Mean   : 6.499   Mean   :200.3   Mean   :  98.95   Mean   : 16.23  
 3rd Qu.: 9.000   3rd Qu.:255.0   3rd Qu.: 148.00   3rd Qu.: 22.00  
 Max.   :12.000   Max.   :403.0   Max.   : 254.00   Max.   :422.00  
     nevada    prof_nieve          longitud        latitud        
 Min.   :0   Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889  
 1st Qu.:0   1st Qu.:  0.0000   1st Qu.:38.28   1st Qu.: -5.6417  
 Median :0   Median :  0.0000   Median :40.82   Median : -3.4500  
 Mean   :0   Mean   :  0.1602   Mean   :39.66   Mean   : -3.4348  
 3rd Qu.:0   3rd Qu.:  0.0000   3rd Qu.:42.08   3rd Qu.:  0.4914  
 Max.   :0   Max.   :194.0000   Max.   :43.57   Max.   :  4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  42.0  
 Median : 247.0  
 Mean   : 417.6  
 3rd Qu.: 656.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax            tmin           precip          nevada 
 Min.   : 1.000   Min.   :86.00   Min.   :-24.0   Min.   : 6.00   Min.   :1  
 1st Qu.: 6.500   1st Qu.:86.00   1st Qu.:-19.5   1st Qu.: 6.00   1st Qu.:1  
 Median :12.000   Median :86.00   Median :-15.0   Median : 6.00   Median :1  
 Mean   : 8.333   Mean   :88.33   Mean   : -7.0   Mean   :10.33   Mean   :1  
 3rd Qu.:12.000   3rd Qu.:89.50   3rd Qu.:  1.5   3rd Qu.:12.50   3rd Qu.:1  
 Max.   :12.000   Max.   :93.00   Max.   : 18.0   Max.   :19.00   Max.   :1  
   prof_nieve        longitud        latitud          altitud     
 Min.   :0.0000   Min.   :40.48   Min.   :-3.450   Min.   :263.0  
 1st Qu.:0.0000   1st Qu.:40.48   1st Qu.:-3.450   1st Qu.:435.6  
 Median :0.0000   Median :40.48   Median :-3.450   Median :608.1  
 Mean   :0.3333   Mean   :40.88   Mean   :-2.644   Mean   :493.1  
 3rd Qu.:0.5000   3rd Qu.:41.08   3rd Qu.:-2.242   3rd Qu.:608.1  
 Max.   :1.0000   Max.   :41.67   Max.   :-1.033   Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt          tmax             tmin            precip           nevada 
 Min.   :1.000   Min.   :-12.00   Min.   :-72.00   Min.   : 19.00   Min.   :0  
 1st Qu.:2.000   1st Qu.:  9.50   1st Qu.:-43.50   1st Qu.: 41.00   1st Qu.:0  
 Median :2.000   Median : 15.00   Median :-34.00   Median : 56.00   Median :0  
 Mean   :2.368   Mean   : 23.21   Mean   :-31.95   Mean   : 65.79   Mean   :0  
 3rd Qu.:3.000   3rd Qu.: 28.50   3rd Qu.:-25.00   3rd Qu.: 82.00   3rd Qu.:0  
 Max.   :4.000   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :465.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:531.0   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :606.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :612.8   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:661.5   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :892.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax             tmin            precip          nevada 
 Min.   : 1.000   Min.   :-12.00   Min.   :-63.00   Min.   : 5.00   Min.   :0  
 1st Qu.: 1.500   1st Qu.: 14.00   1st Qu.:-38.00   1st Qu.:25.50   1st Qu.:0  
 Median : 2.000   Median : 28.00   Median :-26.00   Median :48.00   Median :0  
 Mean   : 3.308   Mean   : 31.95   Mean   :-25.74   Mean   :46.64   Mean   :0  
 3rd Qu.: 3.500   3rd Qu.: 46.50   3rd Qu.:-15.00   3rd Qu.:66.50   3rd Qu.:0  
 Max.   :12.000   Max.   : 96.00   Max.   : 28.00   Max.   :92.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :197.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:227.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :269.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :278.4   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:318.0   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :429.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt        tmax           tmin           precip         nevada 
 Min.   :2.0   Min.   :-4.0   Min.   :-51.0   Min.   :19.0   Min.   :0  
 1st Qu.:2.0   1st Qu.: 0.0   1st Qu.:-48.0   1st Qu.:33.0   1st Qu.:0  
 Median :3.0   Median :35.0   Median :-26.0   Median :49.0   Median :0  
 Mean   :2.6   Mean   :20.6   Mean   :-32.8   Mean   :41.4   Mean   :0  
 3rd Qu.:3.0   3rd Qu.:36.0   3rd Qu.:-23.0   3rd Qu.:50.0   3rd Qu.:0  
 Max.   :3.0   Max.   :36.0   Max.   :-16.0   Max.   :56.0   Max.   :0  
   prof_nieve      longitud        latitud         altitud    
 Min.   :1017   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:1073   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :1168   Median :40.78   Median :-4.01   Median :1894  
 Mean   :1317   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:1494   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster06)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1], dim(df.cluster06)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05", "cluster06"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.hist(df.cluster06)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster06)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
  df.cluster06.grouped <- mpr.group_by_geo(df.cluster06)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster06.grouped)

Visualización de 8 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=8)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=8)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)
  df.cluster06 <- subset(df, cluster==6)
  df.cluster07 <- subset(df, cluster==7)
  df.cluster08 <- subset(df, cluster==8)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster06 <- select(df.cluster06, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster07 <- select(df.cluster07, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster08 <- select(df.cluster08, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax          tmin             precip           nevada 
 Min.   : 1.000   Min.   :-53   Min.   :-121.00   Min.   :  0.00   Min.   :0  
 1st Qu.: 2.000   1st Qu.:122   1st Qu.:  28.00   1st Qu.:  6.00   1st Qu.:0  
 Median : 4.000   Median :156   Median :  59.00   Median : 14.00   Median :0  
 Mean   : 5.912   Mean   :152   Mean   :  56.19   Mean   : 19.76   Mean   :0  
 3rd Qu.:11.000   3rd Qu.:184   3rd Qu.:  86.00   3rd Qu.: 28.00   3rd Qu.:0  
 Max.   :12.000   Max.   :336   Max.   : 219.00   Max.   :126.00   Max.   :0  
   prof_nieve          longitud        latitud            altitud      
 Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.:  0.0000   1st Qu.:39.47   1st Qu.: -5.3456   1st Qu.:  64.0  
 Median :  0.0000   Median :41.11   Median : -2.7331   Median : 412.0  
 Mean   :  0.2758   Mean   :40.41   Mean   : -2.8404   Mean   : 514.8  
 3rd Qu.:  0.0000   3rd Qu.:42.33   3rd Qu.:  0.4942   3rd Qu.: 750.0  
 Max.   :194.0000   Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax            tmin           precip           nevada 
 Min.   : 1.000   Min.   :153.0   Min.   : 54.0   Min.   : 0.000   Min.   :0  
 1st Qu.: 6.000   1st Qu.:234.0   1st Qu.:131.0   1st Qu.: 1.000   1st Qu.:0  
 Median : 7.000   Median :264.0   Median :154.0   Median : 5.000   Median :0  
 Mean   : 7.205   Mean   :265.4   Mean   :155.9   Mean   : 9.267   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:294.0   3rd Qu.:181.0   3rd Qu.:15.000   3rd Qu.:0  
 Max.   :12.000   Max.   :403.0   Max.   :254.0   Max.   :53.000   Max.   :0  
   prof_nieve          longitud        latitud            altitud      
 Min.   : 0.00000   Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.: 0.00000   1st Qu.:37.19   1st Qu.: -6.0556   1st Qu.:  32.0  
 Median : 0.00000   Median :40.07   Median : -3.6781   Median :  87.0  
 Mean   : 0.00272   Mean   :38.63   Mean   : -4.1861   Mean   : 291.3  
 3rd Qu.: 0.00000   3rd Qu.:41.66   3rd Qu.:  0.3664   3rd Qu.: 540.0  
 Max.   :35.00000   Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax            tmin            precip          nevada 
 Min.   : 1.000   Min.   : -8.0   Min.   :-58.00   Min.   : 78.0   Min.   :0  
 1st Qu.:10.000   1st Qu.:127.0   1st Qu.: 64.00   1st Qu.: 91.0   1st Qu.:0  
 Median :11.000   Median :149.0   Median : 84.00   Median :106.0   Median :0  
 Mean   : 9.729   Mean   :156.6   Mean   : 88.43   Mean   :116.6   Mean   :0  
 3rd Qu.:12.000   3rd Qu.:182.0   3rd Qu.:111.00   3rd Qu.:131.0   3rd Qu.:0  
 Max.   :12.000   Max.   :350.0   Max.   :223.00   Max.   :422.0   Max.   :0  
   prof_nieve         longitud        latitud           altitud      
 Min.   : 0.0000   Min.   :27.82   Min.   :-17.889   Min.   :   1.0  
 1st Qu.: 0.0000   1st Qu.:40.80   1st Qu.: -8.411   1st Qu.:  32.0  
 Median : 0.0000   Median :42.43   Median : -5.598   Median : 127.0  
 Mean   : 0.3505   Mean   :41.28   Mean   : -5.054   Mean   : 299.2  
 3rd Qu.: 0.0000   3rd Qu.:43.31   3rd Qu.: -2.039   3rd Qu.: 370.0  
 Max.   :81.0000   Max.   :43.57   Max.   :  4.216   Max.   :2519.0  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax            tmin           precip          nevada 
 Min.   : 1.000   Min.   :86.00   Min.   :-24.0   Min.   : 6.00   Min.   :1  
 1st Qu.: 6.500   1st Qu.:86.00   1st Qu.:-19.5   1st Qu.: 6.00   1st Qu.:1  
 Median :12.000   Median :86.00   Median :-15.0   Median : 6.00   Median :1  
 Mean   : 8.333   Mean   :88.33   Mean   : -7.0   Mean   :10.33   Mean   :1  
 3rd Qu.:12.000   3rd Qu.:89.50   3rd Qu.:  1.5   3rd Qu.:12.50   3rd Qu.:1  
 Max.   :12.000   Max.   :93.00   Max.   : 18.0   Max.   :19.00   Max.   :1  
   prof_nieve        longitud        latitud          altitud     
 Min.   :0.0000   Min.   :40.48   Min.   :-3.450   Min.   :263.0  
 1st Qu.:0.0000   1st Qu.:40.48   1st Qu.:-3.450   1st Qu.:435.6  
 Median :0.0000   Median :40.48   Median :-3.450   Median :608.1  
 Mean   :0.3333   Mean   :40.88   Mean   :-2.644   Mean   :493.1  
 3rd Qu.:0.5000   3rd Qu.:41.08   3rd Qu.:-2.242   3rd Qu.:608.1  
 Max.   :1.0000   Max.   :41.67   Max.   :-1.033   Max.   :608.1  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt          tmax             tmin            precip           nevada 
 Min.   :1.000   Min.   :-12.00   Min.   :-72.00   Min.   : 19.00   Min.   :0  
 1st Qu.:2.000   1st Qu.:  9.50   1st Qu.:-43.50   1st Qu.: 41.00   1st Qu.:0  
 Median :2.000   Median : 15.00   Median :-34.00   Median : 56.00   Median :0  
 Mean   :2.368   Mean   : 23.21   Mean   :-31.95   Mean   : 65.79   Mean   :0  
 3rd Qu.:3.000   3rd Qu.: 28.50   3rd Qu.:-25.00   3rd Qu.: 82.00   3rd Qu.:0  
 Max.   :4.000   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :465.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:531.0   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :606.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :612.8   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:661.5   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :892.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster06)

   fecha_cnt           tmax             tmin            precip          nevada 
 Min.   : 1.000   Min.   :-12.00   Min.   :-63.00   Min.   : 5.00   Min.   :0  
 1st Qu.: 1.500   1st Qu.: 14.00   1st Qu.:-38.00   1st Qu.:25.50   1st Qu.:0  
 Median : 2.000   Median : 28.00   Median :-26.00   Median :48.00   Median :0  
 Mean   : 3.308   Mean   : 31.95   Mean   :-25.74   Mean   :46.64   Mean   :0  
 3rd Qu.: 3.500   3rd Qu.: 46.50   3rd Qu.:-15.00   3rd Qu.:66.50   3rd Qu.:0  
 Max.   :12.000   Max.   : 96.00   Max.   : 28.00   Max.   :92.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :197.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:227.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :269.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :278.4   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:318.0   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :429.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster07)

   fecha_cnt        tmax           tmin           precip         nevada 
 Min.   :2.0   Min.   :-4.0   Min.   :-51.0   Min.   :19.0   Min.   :0  
 1st Qu.:2.0   1st Qu.: 0.0   1st Qu.:-48.0   1st Qu.:33.0   1st Qu.:0  
 Median :3.0   Median :35.0   Median :-26.0   Median :49.0   Median :0  
 Mean   :2.6   Mean   :20.6   Mean   :-32.8   Mean   :41.4   Mean   :0  
 3rd Qu.:3.0   3rd Qu.:36.0   3rd Qu.:-23.0   3rd Qu.:50.0   3rd Qu.:0  
 Max.   :3.0   Max.   :36.0   Max.   :-16.0   Max.   :56.0   Max.   :0  
   prof_nieve      longitud        latitud         altitud    
 Min.   :1017   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:1073   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :1168   Median :40.78   Median :-4.01   Median :1894  
 Mean   :1317   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:1494   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster08)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1], dim(df.cluster06)[1], dim(df.cluster07)[1], dim(df.cluster08)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05", "cluster06", "cluster07", "cluster08"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.hist(df.cluster06)

if (!empty_nodes) mpr.hist(df.cluster07)

if (!empty_nodes) mpr.hist(df.cluster08)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster06)

if (!empty_nodes) mpr.boxplot(df.cluster07)

if (!empty_nodes) mpr.boxplot(df.cluster08)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
  df.cluster06.grouped <- mpr.group_by_geo(df.cluster06)
  df.cluster07.grouped <- mpr.group_by_geo(df.cluster07)
  df.cluster08.grouped <- mpr.group_by_geo(df.cluster08)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster06.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster07.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster08.grouped)

Visualización de 10 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=10)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=10)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)
  df.cluster06 <- subset(df, cluster==6)
  df.cluster07 <- subset(df, cluster==7)
  df.cluster08 <- subset(df, cluster==8)
  df.cluster09 <- subset(df, cluster==9)
  df.cluster10 <- subset(df, cluster==10)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster06 <- select(df.cluster06, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster07 <- select(df.cluster07, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster08 <- select(df.cluster08, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster09 <- select(df.cluster09, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster10 <- select(df.cluster10, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt         tmax            tmin             precip           nevada 
 Min.   :1.00   Min.   :-53.0   Min.   :-121.00   Min.   :  0.00   Min.   :0  
 1st Qu.:2.00   1st Qu.:123.0   1st Qu.:  26.00   1st Qu.:  5.00   1st Qu.:0  
 Median :3.00   Median :155.0   Median :  56.00   Median : 13.00   Median :0  
 Mean   :2.81   Mean   :150.9   Mean   :  51.64   Mean   : 18.52   Mean   :0  
 3rd Qu.:4.00   3rd Qu.:182.0   3rd Qu.:  80.00   3rd Qu.: 26.00   3rd Qu.:0  
 Max.   :8.00   Max.   :278.0   Max.   : 178.00   Max.   :126.00   Max.   :0  
   prof_nieve         longitud        latitud            altitud      
 Min.   : 0.0000   Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.: 0.0000   1st Qu.:39.01   1st Qu.: -5.4981   1st Qu.:  61.0  
 Median : 0.0000   Median :41.11   Median : -2.9056   Median : 370.0  
 Mean   : 0.1288   Mean   :40.36   Mean   : -2.8768   Mean   : 497.8  
 3rd Qu.: 0.0000   3rd Qu.:42.29   3rd Qu.:  0.4914   3rd Qu.: 704.0  
 Max.   :60.0000   Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax            tmin           precip           nevada 
 Min.   : 1.000   Min.   :153.0   Min.   : 54.0   Min.   : 0.000   Min.   :0  
 1st Qu.: 6.000   1st Qu.:234.0   1st Qu.:131.0   1st Qu.: 1.000   1st Qu.:0  
 Median : 7.000   Median :264.0   Median :154.0   Median : 5.000   Median :0  
 Mean   : 7.205   Mean   :265.4   Mean   :155.9   Mean   : 9.267   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:294.0   3rd Qu.:181.0   3rd Qu.:15.000   3rd Qu.:0  
 Max.   :12.000   Max.   :403.0   Max.   :254.0   Max.   :53.000   Max.   :0  
   prof_nieve          longitud        latitud            altitud      
 Min.   : 0.00000   Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.: 0.00000   1st Qu.:37.19   1st Qu.: -6.0556   1st Qu.:  32.0  
 Median : 0.00000   Median :40.07   Median : -3.6781   Median :  87.0  
 Mean   : 0.00272   Mean   :38.63   Mean   : -4.1861   Mean   : 291.3  
 3rd Qu.: 0.00000   3rd Qu.:41.66   3rd Qu.:  0.3664   3rd Qu.: 540.0  
 Max.   :35.00000   Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt          tmax            tmin             precip           nevada 
 Min.   : 5.00   Min.   :-24.0   Min.   :-110.00   Min.   :  0.00   Min.   :0  
 1st Qu.:10.00   1st Qu.:121.0   1st Qu.:  34.00   1st Qu.:  7.00   1st Qu.:0  
 Median :11.00   Median :158.0   Median :  67.00   Median : 16.00   Median :0  
 Mean   :10.96   Mean   :154.1   Mean   :  63.86   Mean   : 21.71   Mean   :0  
 3rd Qu.:12.00   3rd Qu.:189.0   3rd Qu.:  95.00   3rd Qu.: 31.00   3rd Qu.:0  
 Max.   :12.00   Max.   :336.0   Max.   : 219.00   Max.   :122.00   Max.   :0  
   prof_nieve        longitud        latitud            altitud      
 Min.   : 0.000   Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.: 0.000   1st Qu.:39.49   1st Qu.: -5.2892   1st Qu.:  68.6  
 Median : 0.000   Median :41.15   Median : -2.7331   Median : 421.0  
 Mean   : 0.105   Mean   :40.50   Mean   : -2.7771   Mean   : 537.5  
 3rd Qu.: 0.000   3rd Qu.:42.36   3rd Qu.:  0.4942   3rd Qu.: 779.0  
 Max.   :59.000   Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax             tmin            precip      
 Min.   : 1.000   Min.   : -2.00   Min.   :-50.00   Min.   :  0.00  
 1st Qu.: 1.000   1st Qu.: 28.25   1st Qu.:-27.75   1st Qu.: 21.25  
 Median : 3.000   Median : 40.00   Median :-21.00   Median : 37.50  
 Mean   : 5.284   Mean   : 44.77   Mean   :-18.01   Mean   : 40.65  
 3rd Qu.:11.000   3rd Qu.: 59.00   3rd Qu.: -6.25   3rd Qu.: 53.75  
 Max.   :12.000   Max.   :119.00   Max.   : 30.00   Max.   :126.00  
     nevada    prof_nieve        longitud        latitud          altitud    
 Min.   :0   Min.   : 62.00   Min.   :40.35   Min.   :-4.680   Min.   : 900  
 1st Qu.:0   1st Qu.: 81.25   1st Qu.:40.78   1st Qu.:-4.010   1st Qu.:1894  
 Median :0   Median :103.50   Median :40.78   Median :-4.010   Median :1894  
 Mean   :0   Mean   :112.96   Mean   :40.79   Mean   :-3.960   Mean   :1859  
 3rd Qu.:0   3rd Qu.:141.75   3rd Qu.:40.78   3rd Qu.:-4.010   3rd Qu.:1894  
 Max.   :0   Max.   :194.00   Max.   :41.77   Max.   :-1.117   Max.   :1894  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt           tmax            tmin            precip          nevada 
 Min.   : 1.000   Min.   : -8.0   Min.   :-58.00   Min.   : 78.0   Min.   :0  
 1st Qu.:10.000   1st Qu.:127.0   1st Qu.: 64.00   1st Qu.: 91.0   1st Qu.:0  
 Median :11.000   Median :149.0   Median : 84.00   Median :106.0   Median :0  
 Mean   : 9.729   Mean   :156.6   Mean   : 88.43   Mean   :116.6   Mean   :0  
 3rd Qu.:12.000   3rd Qu.:182.0   3rd Qu.:111.00   3rd Qu.:131.0   3rd Qu.:0  
 Max.   :12.000   Max.   :350.0   Max.   :223.00   Max.   :422.0   Max.   :0  
   prof_nieve         longitud        latitud           altitud      
 Min.   : 0.0000   Min.   :27.82   Min.   :-17.889   Min.   :   1.0  
 1st Qu.: 0.0000   1st Qu.:40.80   1st Qu.: -8.411   1st Qu.:  32.0  
 Median : 0.0000   Median :42.43   Median : -5.598   Median : 127.0  
 Mean   : 0.3505   Mean   :41.28   Mean   : -5.054   Mean   : 299.2  
 3rd Qu.: 0.0000   3rd Qu.:43.31   3rd Qu.: -2.039   3rd Qu.: 370.0  
 Max.   :81.0000   Max.   :43.57   Max.   :  4.216   Max.   :2519.0  

if (!empty_nodes) summary(df.cluster06)

   fecha_cnt           tmax            tmin           precip          nevada 
 Min.   : 1.000   Min.   :86.00   Min.   :-24.0   Min.   : 6.00   Min.   :1  
 1st Qu.: 6.500   1st Qu.:86.00   1st Qu.:-19.5   1st Qu.: 6.00   1st Qu.:1  
 Median :12.000   Median :86.00   Median :-15.0   Median : 6.00   Median :1  
 Mean   : 8.333   Mean   :88.33   Mean   : -7.0   Mean   :10.33   Mean   :1  
 3rd Qu.:12.000   3rd Qu.:89.50   3rd Qu.:  1.5   3rd Qu.:12.50   3rd Qu.:1  
 Max.   :12.000   Max.   :93.00   Max.   : 18.0   Max.   :19.00   Max.   :1  
   prof_nieve        longitud        latitud          altitud     
 Min.   :0.0000   Min.   :40.48   Min.   :-3.450   Min.   :263.0  
 1st Qu.:0.0000   1st Qu.:40.48   1st Qu.:-3.450   1st Qu.:435.6  
 Median :0.0000   Median :40.48   Median :-3.450   Median :608.1  
 Mean   :0.3333   Mean   :40.88   Mean   :-2.644   Mean   :493.1  
 3rd Qu.:0.5000   3rd Qu.:41.08   3rd Qu.:-2.242   3rd Qu.:608.1  
 Max.   :1.0000   Max.   :41.67   Max.   :-1.033   Max.   :608.1  

if (!empty_nodes) summary(df.cluster07)

   fecha_cnt          tmax             tmin            precip           nevada 
 Min.   :1.000   Min.   :-12.00   Min.   :-72.00   Min.   : 19.00   Min.   :0  
 1st Qu.:2.000   1st Qu.:  9.50   1st Qu.:-43.50   1st Qu.: 41.00   1st Qu.:0  
 Median :2.000   Median : 15.00   Median :-34.00   Median : 56.00   Median :0  
 Mean   :2.368   Mean   : 23.21   Mean   :-31.95   Mean   : 65.79   Mean   :0  
 3rd Qu.:3.000   3rd Qu.: 28.50   3rd Qu.:-25.00   3rd Qu.: 82.00   3rd Qu.:0  
 Max.   :4.000   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :465.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:531.0   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :606.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :612.8   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:661.5   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :892.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster08)

   fecha_cnt           tmax             tmin            precip          nevada 
 Min.   : 1.000   Min.   :-12.00   Min.   :-63.00   Min.   : 5.00   Min.   :0  
 1st Qu.: 1.500   1st Qu.: 14.00   1st Qu.:-38.00   1st Qu.:25.50   1st Qu.:0  
 Median : 2.000   Median : 28.00   Median :-26.00   Median :48.00   Median :0  
 Mean   : 3.308   Mean   : 31.95   Mean   :-25.74   Mean   :46.64   Mean   :0  
 3rd Qu.: 3.500   3rd Qu.: 46.50   3rd Qu.:-15.00   3rd Qu.:66.50   3rd Qu.:0  
 Max.   :12.000   Max.   : 96.00   Max.   : 28.00   Max.   :92.00   Max.   :0  
   prof_nieve       longitud        latitud         altitud    
 Min.   :197.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:227.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :269.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   :278.4   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:318.0   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :429.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster09)

   fecha_cnt        tmax           tmin           precip         nevada 
 Min.   :2.0   Min.   :-4.0   Min.   :-51.0   Min.   :19.0   Min.   :0  
 1st Qu.:2.0   1st Qu.: 0.0   1st Qu.:-48.0   1st Qu.:33.0   1st Qu.:0  
 Median :3.0   Median :35.0   Median :-26.0   Median :49.0   Median :0  
 Mean   :2.6   Mean   :20.6   Mean   :-32.8   Mean   :41.4   Mean   :0  
 3rd Qu.:3.0   3rd Qu.:36.0   3rd Qu.:-23.0   3rd Qu.:50.0   3rd Qu.:0  
 Max.   :3.0   Max.   :36.0   Max.   :-16.0   Max.   :56.0   Max.   :0  
   prof_nieve      longitud        latitud         altitud    
 Min.   :1017   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.:1073   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median :1168   Median :40.78   Median :-4.01   Median :1894  
 Mean   :1317   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.:1494   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834   Max.   :40.78   Max.   :-4.01   Max.   :1894  

if (!empty_nodes) summary(df.cluster10)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1], dim(df.cluster06)[1], dim(df.cluster07)[1], dim(df.cluster08)[1], dim(df.cluster09)[1], dim(df.cluster10)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05", "cluster06", "cluster07", "cluster08", "cluster09", "cluster10"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.hist(df.cluster06)

if (!empty_nodes) mpr.hist(df.cluster07)

if (!empty_nodes) mpr.hist(df.cluster08)

if (!empty_nodes) mpr.hist(df.cluster09)

if (!empty_nodes) mpr.hist(df.cluster10)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster06)

if (!empty_nodes) mpr.boxplot(df.cluster07)

if (!empty_nodes) mpr.boxplot(df.cluster08)

if (!empty_nodes) mpr.boxplot(df.cluster09)

if (!empty_nodes) mpr.boxplot(df.cluster10)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
  df.cluster06.grouped <- mpr.group_by_geo(df.cluster06)
  df.cluster07.grouped <- mpr.group_by_geo(df.cluster07)
  df.cluster08.grouped <- mpr.group_by_geo(df.cluster08)
  df.cluster09.grouped <- mpr.group_by_geo(df.cluster09)
  df.cluster10.grouped <- mpr.group_by_geo(df.cluster10)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster06.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster07.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster08.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster09.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster10.grouped)